The function will take the following arguments:

• myScatterInput: a data frame with n rows and m columns, where all entries will be real numbers

• myClusterNum: an integer value greater than or equal to 2 but less than or equal to n – 2 < myClusterNum < n

The funciton will take the following steps:

1. Randomly assign each of the points in myScatterInput to some value between 1 and myClusterNum.

2. Compute the cluster centroid for each cluster 1 to myClusterNum.

3. Compute the Euclidean distance from each cluster centroid to each data point.

4. Assign each point to the cluster centroid which minimizes the Euclidean distance.

5. Repeat steps 2, 3 and 4 until one of two stopping conditions are met

   • subsequent cluster assignments are unchanged

   • have repeated steps 2, 3 and 4 maxIter number of times

6. Once reaching a terminating condition, compute the sum of all Euclidean distances from each point to their respective centroids.

7. Repeat steps 1-6 100 times.

8. Identify the replication with the lowest sum of Euclidean distances from points to centroids as best result and print the value to the console.